BACKGROUND OF THE INVENTION
[0001] The present invention relates to a snapshot creating method, a storage system, and
a disk apparatus. In particular, it relates to a snapshot creating method, a disk
apparatus, and a storage system in a network-attached storage system for performing
a file service via a network.
[0002] U.S. Patents No. 5,790,773, No. 5,819,292, and No. 5,835,953 have disclosed a computer
file system, a snapshot generation, and a mirror disk component. As a prior art on
a storage system, there has been proposed and known the network-attached storage (NAS).
Based on the NAS, a special-purpose server for providing a file service to a disk
apparatus via a network becomes capable of accessing data within the disk apparatus
from the network.
[0003] Generally speaking, data stored in a storage apparatus necessitates the execution
of the back-up processing in preparation for the case where the data has been lost
by a malfunction in the apparatus, a defect in a software program, an erroneous operation,
or the like. Here, the back-up processing is a processing for copying and saving the
data regularly into a magnetic tape or the like so that the lost data will be able
to be recovered. The back-up processing is similarly needed in the NAS as well.
[0004] If, during the back-up processing task, the data recorded in the storage apparatus
has been updated, there occurs an inconsistency in the copied data. Accordingly, it
is necessary to prevent the data from being updated during the back-up task so as
to guarantee a stored image of the data at a back-up starting point-in-time. In order
to avoid the updating of the data to be backed up, it is all right to shutdown all
of the programs other than the back-up program which are going to access the data.
In the case of a high-availability requested system, however, it is impossible to
shutdown the programs for a long time. This situation requires a mechanism for not
preventing the programs from updating the data during the back-up, and at the same
time for saving the stored image of the data at the back-up starting point-in-time.
Usually, a stored image, or a temporal copy, of data at a point-in-time is referred
to as "snapshot", and the data that has becomes the target of the snapshot creation
is referred to as "operational data". In the NAS, as methods of providing a state
where it is possible to update the data while saving the snapshot at a specified point-in-time,
the following methods have been known: Duplexing of the data by the disk apparatus,
saving of the pre-updated data by the NAS processing unit, and the like.
[0005] Next, the explanation will be given below concerning the snapshot creating methods
according to the above-described prior arts.
[0006] As the prior art on the duplexing of the data by the disk apparatus, there has been
known the technology described in the above-described U.S. Patent No. 5,790,773. In
this prior art, from a normal state before a snapshot creating instruction is issued,
all the data within the disk apparatus are duplexed (i.e., mirrored) into a second
storage region. Moreover, in this prior art, when creating a snapshot, a management
program within the disk apparatus provides one region for the operational data, and
another region for the snapshot data.
[0007] Meanwhile, the saving of the pre-updated data by the NAS processing unit is performed
by saving the pre-updated stored contents into a different storage region at the time
of a data updating after a snapshot creating instruction has been issued from the
outside. Logically, the snapshot is accessed as different data that is independent
of the operational data. However, a part of the snapshot whose operational data has
been not updated after the snapshot creating instruction shares its storage region
with the operational data. Only a part of the snapshot whose operational data has
been updated is created by making reference to the pre-updated stored contents saved
into the different storage region. In the case of the NAS, there exists a snapshot
technology in the file or directory unit. Namely, when a snapshot creating instruction
is issued from the outside, the NAS processing unit, if a file updating occurs thereinafter,
saves the pre-updated file image into a different storage region so as to prevent
the data crash from occurring. In this way, the snapshot technology allows the execution
of the snapshot creation.
[0008] In the above-described prior art regarding the snapshot creating method based on
the data duplexing processing by the disk apparatus, the duplexing processing must
be executed all over the entire physical volumes. Also, a physical volume becomes
necessary whose size is the same as that of the physical volume into which the operational
data has been stored. This results in a problem that the necessary storage capacity
becomes 2 times larger. Also, in this prior art, when the disk apparatus performs
a snapshot creation, the disk apparatus cannot differentiate the data contents of
the respective blocks within the physical volume. As a result, the data duplexing
is executed with respect to the case of a data block as well which is meaningless
to the NAS processing unit, the examples of which are: The case where the data contents
within the block have been formatted by the NAS processing unit, the case where, although
there exists the data within the block, the data has been not used by the NAS processing
unit, and the like. This gives rise to a problem of increasing a CPU load in the disk
apparatus, a data communications amount within the disk apparatus, and the like.
[0009] Also, the above-described prior art regarding the snapshot creating method by the
NAS processing unit is carried out under a condition that the NAS processing unit
performs all of the saving/management of the pre-updated data. This condition increases
a CPU load in the NAS processing unit, and a data communications amount in the network
for connecting the NAS processing unit to the disk apparatus. As a result, there exists
a problem that the execution speed of an application program is exceedingly lowered
in comparison with the one in a snapshot non-creating system. In particular, there
occurs a problem that the performance reduction becomes conspicuous in a processing
accompanied by a large quantity of data updating, e.g., a replication of the database.
SUMMARY OF THE INVENTION
[0010] It is an object of the present invention to solve the above-described problems, and
thereby to provide a snapshot creating method, a disk apparatus and a storage system
using this method. Here, in the snapshot creation by the disk apparatus, the snapshot
creating method allows the issuing of a snapshot creating instruction in a partition
unit.
[0011] Also, it is another object of the present invention to provide a snapshot creating
method, a disk apparatus and a storage system using this method. Here, in the snapshot
creation by the disk apparatus, the snapshot creating method reduces a data communications
amount and a necessary storage capacity within the disk apparatus.
[0012] It is still another object of the present invention to provide a snapshot creating
method, a disk apparatus and a storage system using this method. Here, in the snapshot
creation by the disk apparatus, the snapshot creating method allows a snapshot creating
instruction in a file/directory unit, decreases a CPU load in the NAS processing unit,
and allows reductions in a data processing amount and a necessary storage capacity
within the disk apparatus.
[0013] According to the present invention, there is provided a storage system, comprising:
a 1st data storage, a 2nd data storage connected to said 1st data storage, a processor
for controlling said 1st data storage and said 2nd data storage, and data-block analyzing
units for receiving data-block information so as to store said data-block information
into a storage region, said data-block information indicating which data block is
a useful data block out of a plurality of data blocks in said 1st data storage.
[0014] According to the present invention, the above-described objects are accomplished
as follows: In a snapshot creating method in a storage system including a file server
and a disk apparatus, there are provided the following steps: A step of the file server's
instructing the disk apparatus in a block whose snapshot data should be generated,
and if there occurs an updating request for operational data, a step of the disk apparatus's
confirming whether or not an updated-data written-in destination block in an operational
data storing region has become a snapshot-data creation specified block, and if the
updated-data written-in destination block has not become a target of the snapshot
data generation, a step of the disk apparatus's instructing a disk controller to write
updated data into the operational data storing region, and if the updated-data written-in
destination block has become the snapshot-data generation specified block, a step
of the disk apparatus's confirming whether or not the specified block had been copied
into a snapshot data storing region before, and if the specified block had been copied
before, a step of the disk apparatus's instructing the disk controller to write the
updated data into the operational data storing region, and if the specified block
had not been copied before, a step of the disk apparatus's instructing the disk controller
to copy pre-updated data into the snapshot data storing region, the pre-updated data
existing in the operational data storing region, and a step of the disk apparatus's
receiving the instruction of copying the pre-updated data, and confirming whether
or not the snapshot data storing region has been ensured, and if the snapshot data
storing region has not been ensured, a step of the disk apparatus's ensuring the snapshot
data storing region, and a step of the disk apparatus's copying the pre-updated data
into the snapshot data storing region, and at the same time generating a block-address
correspondence table if necessary, and a step of the disk apparatus's writing the
updated data into the written-in destination block in the operational data storing
region.
[0015] Also, the above-described objects are accomplished by configuring a disk apparatus
as follows: The disk apparatus includes a specified-block analyzing unit for receiving,
from a file server, an instruction of a block specified as the target of a snapshot
creation, and for making a judgement as to whether the snapshot data generation is
of a necessity or not in accordance with the instruction, a disk controller for executing
the data copy on the basis of the judgement by the specified-block analyzing unit,
an operational data storing region for saving operational data, and a snapshot data
storing region for saving the snapshot data.
[0016] Moreover, the above-described objects are accomplished as follows: In a storage system
including a file server and a disk apparatus, the file server includes an instruction
unit for instructing the disk apparatus in a block whose snapshot data should be generated,
and a transmission unit for transmitting updated data to the disk apparatus, and at
the same time transmitting the instruction of the block as additional information.
BRIEF DESCRIPTION OF THE DRAWINGS
[0017]
FIG. 1 is a block diagram for illustrating the configuration of a storage system according
to a 1st embodiment of the present invention;
FIG. 2 is a block diagram for illustrating the hardware configuration of the storage
system according to the 1st embodiment of the present invention illustrated in FIG.
1;
FIG. 3 is a diagram for illustrating a configuration example of a block-selection
specifying unit within a NAS processing unit including a specified-block managing
unit for managing blocks that become necessary for the snapshot creation of a partition
specified from the outside;
FIG. 4 is a diagram for illustrating a configuration example of a specified-block
analyzing unit within a disk apparatus including a specified-block managing unit for
managing blocks that become necessary for the snapshot creation of a partition specified
from the outside;
FIG. 5 is a diagram for illustrating an example of a block-address correspondence
table between an operational data storing region and a snapshot data storing region;
FIG. 6 is a flowchart for explaining the processing operation in the case of the snapshot
creation based on the saving of pre-updated data in the disk apparatus;
FIG. 7 is a diagram for explaining a layout example of a file system in the disk apparatus;
FIG. 8 is a diagram for explaining an example of the logical structure of the file
system;
FIG. 9 is a flowchart for explaining the processing steps of the block selection performed
by the block-selection specifying unit in the case where a specified target of the
snapshot data generation specified by a user is a file; and
FIG. 10 is a flowchart for explaining the processing steps of the block selection
performed by the block-selection specifying unit in the case where the specified target
of the snapshot data generation specified by the user is a directory.
DESCRIPTION OF THE EMBODIMENTS
[0018] Hereinafter, referring to the drawings, the detailed explanation will be given below
concerning a snapshot creating method according to the present invention, a storage
system and a disk apparatus using this method.
[0019] FIG. 1 is a block diagram for illustrating the configuration of a storage system
according to a 1st embodiment of the present invention. In FIG. 1, the reference numerals
denote the following configuration components: 1 a disk apparatus, 2 a NAS processing
unit, 3 a network, 11 a specified-block analyzing unit, 12 a disk controller, 13 an
operational data storing region, 14 a snapshot data storing region, 15, 22 interfaces,
21 a disk-block managing unit, 131 to 136, 141 to 143 data blocks, 211 a block-selection
specifying unit.
[0020] As illustrated in FIG. 1, the storage system according to the 1st embodiment of the
present invention includes the following: The disk apparatus 1 for saving data, the
NAS processing unit 2 that, as a file server, performs a processing for providing
a file service via the network, and the network 3 for connecting the disk apparatus
1 to the NAS processing unit 2.
[0021] In addition, the disk apparatus 1 includes the following: The specified-block analyzing
unit 11 for performing the analysis of a specified block and the management of snapshot
data, the disk controller 12 for performing the allocation of a physical volume, the
copy of block data, and the writing-in of updated data, the operational data storing
region 13 for saving operational data, the snapshot data storing region 14 for saving
the snapshot data, and the interface 15 for performing the communications with the
NAS processing unit 2. The operational data storing region 13 and the snapshot data
storing region 14 store the data blocks 131 to 136 and the ones 141 to 143, respectively.
[0022] The NAS processing unit 2 includes the disk-block managing unit 21 as a partition
managing unit for performing the management of a partition, and the interface 22 for
performing the communications with the disk apparatus 1. Namely, the disk-block managing
unit 21, in the 1st embodiment, has a function as the partition managing unit, and
includes the block-selection specifying unit 211 inside. The block-selection specifying
unit 211 selects a block necessary for generating the snapshot data, then specifying
the selected block to the specified-block analyzing unit 11 within the disk apparatus
1.
[0023] In the 1st embodiment of the present invention configured as described above, consider
the case where a snapshot-data generating instruction with respect to one partition
or plural partitions (i.e., logical volume or volumes) within a physical volume is
issued from the outside to the NAS processing unit 2. In this case, the block-selection
specifying unit 211 selects all the blocks included within the above-described one
partition or plural partitions, then specifying these selected blocks to the specified-block
analyzing unit 11 within the disk apparatus 1.
[0024] FIG. 2 is a block diagram for illustrating the hardware configuration of the storage
system according to the 1st embodiment of the present invention illustrated in FIG.
1. In FIG. 2, the reference numerals denote the following configuration components:
16, 17 interfaces, 18, 25 CPUs, 19, 26 memories, 11' a specified-block analyzing program,
12' a disk controlling program, 21' a disk-block managing program, 211' a block-selection
instructing module. The other reference numerals are the same as those in FIG. 1.
Although the interfaces 22 and 15 are connected to the interface 16 or 17 in series,
the number of the interfaces can be decreased in the system design.
[0025] As illustrated in FIG. 2, the NAS processing unit 2 includes the CPU 25 and the memory
26. Moreover, for the CPU 25 to execute the disk-block managing program 21' within
the memory 26 implements the disk-block managing unit 21 illustrated in FIG. 1. Also,
for the CPU 25 to execute the block-selection instructing module 211', which is a
part of the disk-block managing program 21', implements the block-selection specifying
unit 211.
[0026] The disk apparatus 1, in addition to the operational data storing region 13 and the
snapshot data storing region 14, includes the CPU 18, the memory 19, and the interfaces
16 and 17 corresponding to the operational data storing region 13 and the snapshot
data storing region 14 respectively. Furthermore, for the CPU 18 to execute the disk
controlling program 12' within the memory 19 implements the disk controller 12 illustrated
in FIG. 1. Also, for the CPU 18 to execute the specified-block analyzing program 11'
within the memory 19 implements the specified-block analyzing unit 11.
[0027] Incidentally, in the embodiment illustrated in FIG. 2, the configuration has been
employed where the one set of CPU 18 executes both of the programs, i.e., the disk
controlling program 12' and the specified-block analyzing program 11'. It is also
possible, however, to provide a plurality of CPUs in the disk apparatus 1 so that
the disk controlling program 12' and the specified-block analyzing program 11' are
caused to be executed by the different CPUs respectively.
[0028] As the configuration of a snapshot target-block specifying unit for instructing the
snapshot target blocks from the block-selection specifying unit 211 to the specified-block
analyzing unit 11 as was explained earlier, there exist the following two examples:
Next, the explanation will be given regarding these examples.
[0029] FIG. 3 is about an example where a specified-block managing unit 2111 for managing
the target blocks that become necessary for the snapshot creation of the partition
or partitions specified from the outside is provided within the block-selection specifying
unit 211 within the NAS processing unit 2.
[0030] In the case of the example illustrated in FIG. 3, the specification of the target
blocks from the block-selection specifying unit 211 to the specified-block analyzing
unit 11 is carried out as follows: The specification is attached to a command to be
transmitted from the NAS processing unit 2 to the disk apparatus 1 via the interfaces
22 and 15, or, a new command is added to the block-selection specifying unit 211.
Having received these commands, the specified-block analyzing unit 11 (FIG. 1) performs
the analysis of the information on the target-block specification attached to the
command, or makes an inquiry using the command newly added to the block-selection
specifying unit 211. These tasks allow the specified-block analyzing unit 11 to confirm
whether or not, as will be described later, the target blocks are snapshot-data generation
target-blocks. If the target blocks are the snapshot-data generation target-blocks,
the specified-block analyzing unit 11 instructs the disk controller 12 to generate
the snapshot data into the snapshot data storing region 14. The specified-block managing
unit can be provided within a data processing unit (not illustrated) in a disk subsystem
that has the other type of processing unit instead of the NAS processing unit 2 illustrated
in FIG. 1 or FIG. 2.
[0031] FIG. 4 is about an example where a specified-block managing unit 111 for managing
the target blocks that become necessary for the snapshot creation of the partition
or partitions specified from the outside is provided within the specified-block analyzing
unit 11 within the disk apparatus 1.
[0032] In the case of the example illustrated in FIG. 4, the specification of the target
blocks from the block-selection specifying unit 211 (FIG. 1) to the specified-block
analyzing unit 11 (FIG. 4) is carried out as follows: At a point-in-time when the
block-selection specifying unit 211 has completed the selection of all the blocks
included in the partition or the partitions specified, the block-selection specifying
unit 211 writes a block bit map into a block bit map storing region 112 (FIG. 4) provided
at a determined position within the disk apparatus 1. Here, the block bit map indicates
whether or not the respective blocks within the physical volume are the snapshot-data
generation target-blocks. When the snapshot-data generating instruction is issued
from the NAS processing unit 2, the specified-block analyzing unit 11 makes reference
to this block bit map, thereby confirming whether or not the target blocks are the
snapshot-data generation specified-blocks. If the target blocks are the snapshot-data
generation specified-blocks, the specified-block analyzing unit 11 causes the disk
controller 12 to generate the snapshot data into the snapshot data storing region
14.
[0033] In the above explanation of the embodiments of FIGS. 3 and 4, blocks in partitions
are specified, however, it is possible to specify blocks including data available
for the NAS processing unit, or blocks data of file directories specified for creating
the snapshot from outside as explained in the second and third embodiments later.
[0034] FIG. 5 is a diagram for illustrating an example of a block-address correspondence
table 400 between the operational data storing region and the snapshot data storing
region.
[0035] The disk controller 12 manages the block-address correspondence table 400 in order
to manage the correspondences between storing-region identifiers (i.e., storing-region
identifying numbers) and block addresses in the operational data storing region, and
storing-region identifiers (i.e., storing-region identifying numbers) and block addresses
in the snapshot data storing region. From the user, reference can be made to the snapshot
data as one physical volume.
[0036] It is required to indicate that the block addresses of the snapshot data are the
same as those of the operational data. Consequently, when generating the snapshot
data, the disk controller 12 generates and manages the block-address correspondence
table 400, i.e., the correspondence table between the block addresses in the operational
data storing region and the ones in the snapshot data storing region.
[0037] If the user makes reference to the snapshot data, the disk controller 12 causes the
user to make reference thereto at virtual block addresses that are the same as the
block addresses of the operational data. If the user makes reference to the respective
block addresses, the disk controller 12, based on the block-address correspondence
table 400, converts the virtual block addresses into the block addresses in the snapshot
data storing region, thereby making reference to the snapshot data corresponding to
the respective block addresses of the operational data.
[0038] Consider the case where, using the above-described method, a data duplexing instruction
is issued from the NAS processing unit 2 to the disk apparatus 1 in order to perform
the snapshot creation based on the data duplexing by the disk apparatus 1. In this
case, the disk controller 12 ensures a physical volume needed for the snapshot data
generation. Moreover, the disk controller 12 judges whether or not the respective
blocks are the blocks described in the correspondence table, thereby confirming whether
or not the respective blocks have become the snapshot target blocks. Furthermore,
the disk controller 12 performs the duplexing concerning only the blocks that have
become the target blocks. At this time, the block addresses are required to point
at the same data. Accordingly, every time, in the duplexing processing, the disk controller
12 performs the copy and the synchronized writing-in of the respective blocks, the
disk controller 12 generates the block-address correspondence table 400 explained
using FIG. 5. The storage capacity needed for the duplexing may be of the same size
as that of the partition or partitions having become the snapshot creation target.
This makes it possible to reduce the storage capacity needed for the snapshot data
generation.
[0039] FIG. 6 is a flowchart for explaining the processing operation in the case of the
snapshot creation based on the saving of pre-updated data in the disk apparatus 1.
Next, the explanation will be given regarding this flowchart.
(1) If there occurs an updating request for operational data, the specified-block
analyzing unit 11 makes reference to the correspondence table 400 and thereby to confirm
whether or not an updated-data written-in destination block in the operational data
storing region 13 has become a snapshot-data creation specified block (step 300).
(2) If, in the judgement at the step 300, the updated-data written-in destination
block has not become a target of the snapshot data generation, the specified-block
analyzing unit 11 instructs the disk controller 12 to write the updated data into
the operational data storing region 13, then terminating the processing (step 360).
(3) If, in the judgement at the step 300, the updated-data written-in destination
block has become the snapshot-data generation specified block, the specified-block
analyzing unit 11 confirms whether or not the specified block had been copied into
the snapshot data storing region 14 before (step 310).
(4) If, in the judgement at the step 310, the specified block had been copied before,
the analyzing unit 11 instructs the disk controller 12 to write the updated data into
the operational data storing region 13, then terminating the processing (step 360).
(5) If, in the judgement at the step 310, the specified block had not been copied
before, the analyzing unit 11 instructs the disk controller 12 to copy pre-updated
data into the snapshot data storing region 14, the pre-updated data existing in the
operational data storing region 13 (step 320).
(6) Having received the instruction of copying the pre-updated data, the disk controller
12 confirms whether or not the snapshot data storing region 14 has been ensured (step
330).
(7) If, in the judgement at the step 330, the snapshot data storing region 14 has
not been ensured, the disk controller 12 ensures the snapshot data storing region
14 (step 340).
(8) After having ensured the snapshot data storing region 14 in the processing at
the step 340, the disk controller 12 copies the pre-updated data into the snapshot
data storing region 14, and at the same time generating the block-address correspondence
table if necessary (step 350).
(9) Also, if, in the judgement at the step 330, the snapshot data storing region 14
has been ensured, the disk controller 12 copies the pre-updated data into the snapshot
data storing region 14, and at the same time generating the block-address correspondence
table if necessary, the pre-updated data existing in the operational data storing
region 13 (step 350).
(10) The disk controller 12 writes the updated data into the updated-data written-in
destination block in the operational data storing region 13, then terminating the
processing (step 360).
[0040] If reference is made from the user to the physical volume of the snapshot data, the
disk controller 12, when there exists the block-address correspondence table 400,
converts the block addresses on the basis of this block-address correspondence table
400. This permits the user to watch the snapshot data at the block addresses that
are the same as those of the operational data. Also, when there exists the block-address
correspondence table 400, if reference is made to a block of the block address not
described in the block-address correspondence table 400, the disk controller 12 directly
makes reference to the blocks of the operational data.
[0041] The above-described 1st embodiment of the present invention specifies all the blocks
included within the partition or partitions becoming the snapshot creation target.
This allows the specification of the snapshot creation in the partition unit, thus
making it possible to reduce the storage region that becomes necessary for the snapshot
creation.
[0042] Also, in the snapshot creating method based on the data duplexing, the storage region
that becomes necessary for the snapshot creation can be reduced down to a storage
region whose size is the same as the partition size. Still also, in the snapshot creating
method based on the saving of the pre-updated data, the storage region can be reduced
down to a storage region whose size is smaller than the partition size.
[0043] Next, the explanation will be given below concerning a 2nd embodiment of the present
invention. The 2nd embodiment of the present invention is as follows: In the snapshot
creating method based on the data duplexing or the saving of the pre-updated data,
at first, consider a file-system managing unit which corresponds to the disk-block
managing unit 21 provided in the NAS processing unit 2. Next, the block-selection
specifying unit 211 selects a no-data-containing block formatted by the file-system
managing unit, or an in-no-use block whose file has been deleted by the file-system
managing unit. Moreover, the block-selection specifying unit 211 specifies, to the
specified-block analyzing unit 11, the other blocks which include data that are effective
to the NAS processing unit 2. The disk apparatus 1 performs the data duplexing or
the saving of the pre-updated data regarding only these specified blocks, thereby
making it possible to reduce the data processing amount within the disk apparatus
1, and the storage capacity needed for the snapshot creation.
[0044] In the 2nd embodiment of the present invention, the disk-block managing unit 21 illustrated
in FIG. 1 performs the function as the file-system managing unit. Also, in the 2nd
embodiment of the present invention, in the configurations in FIG. 3 and FIG. 4 as
well for illustrating the configuration examples of the target-block specifying unit
from the block-selection specifying unit 211 to the specified-block analyzing unit
11, the disk-block managing unit 21 performs the function as the file-system managing
unit.
[0045] The difference between the 2nd embodiment and the 1st embodiment is as follows: The
target blocks that the block-selection specifying unit 211 selects as the criterion
of the block selection are the blocks including the data that are effective in the
NAS processing unit 2.
[0046] In the 2nd embodiment of the present invention, when performing the snapshot-data
creation based on the data duplexing by the disk apparatus 1, the data duplexing is
performed using a method that is basically the same as the one in the 1st embodiment.
The 2nd embodiment, however, is based on a premise that a file system manages the
data blocks within the physical volume. Thus, the block addresses of the data blocks
are specified in metadata in each file/directory. As a result, if, concerning the
block address of a block where certain data exists, the block address in the operational
data storing region 13 differs from the one in the snapshot data storing region 14,
a consistency as the file system cannot be maintained. This situation, in the 2nd
embodiment of the present invention, makes it absolutely necessary to generate the
block-address correspondence table 400 when performing the saving of the pre-updated
data and the synchronized writing-in of the data in the data duplexing processing.
[0047] Furthermore, in the 2nd embodiment of the present invention, when performing the
snapshot creation based on the saving of the pre-updated data in the disk apparatus
1, a processing is performed which is basically the same as the one in the 1st embodiment
in accordance with the flowchart explained using FIG. 6. In the 2nd embodiment, however,
generating the block-address correspondence table 400 is absolutely necessary in the
processing at the step 350.
[0048] As explained above, according to the 2nd embodiment of the present invention, in
the snapshot-data creation based on the data duplexing or the saving of the pre-updated
data, the disk controller 12 copies only the specified blocks. This makes it possible
to copy only the blocks that have the meaning in the NAS processing unit 2, thereby
allowing the data communications amount within the disk apparatus 1 to be suppressed
down to the smallest possible degree.
[0049] Also, in the above-described 2nd embodiment of the present invention, with respect
to the already deleted blocks specified from the block-selection specifying unit 211
in the NAS processing unit 2 to the specified-block analyzing unit 11, the disk controller
12 clears the data in a block where the data remains. This eliminates the block which
has not been used in the NAS processing unit 2 but in which the data contents remain.
As a result, it also becomes possible to use the block for an enhancement in security.
[0050] Also, the above-described 1st embodiment and 2nd embodiment can also be carried out
in a manner of being co-used together. The co-use allows the snapshot-data generation
to be performed regarding only the data blocks that have the meaning in the NAS processing
unit 2, while simultaneously issuing the snapshot-data generating instruction in the
partition unit. This permits the implementations of a reduction in the storage capacity
at the time of the snapshot-data generation, and of a reduction in the data communications
amount within the disk apparatus 1.
[0051] Next, the explanation will be given below concerning a 3rd embodiment of the present
invention. The 3rd embodiment of the present invention is as follows: In the snapshot
creating method based on the saving of the pre-updated data, concerning a file/directory
specified as a snapshot creation target from the outside by the user, the block-selection
specifying unit 211 selects blocks where there exists data of the file/directory and
blocks where there exists data related therewith, then specifying the selected blocks
to the specified-block managing unit 2111. Moreover, within the disk apparatus 1,
the saving of the pre-updated data is performed regarding only the specified blocks.
This reduces the data processing amount within the disk apparatus 1 and the storage
capacity needed for the snapshot creation, thus decreasing a CPU load for the snapshot
creation in the NAS processing unit 2.
[0052] In the 3rd embodiment of the present invention, the disk-block managing unit 21 illustrated
in FIG. 1 is provided with the function as the file-system managing unit. Also, in
the 3rd embodiment of the present invention, in the configurations in FIG. 3 and FIG.
4 as well for illustrating the configurations of the target-block specifying unit
from the block-selection specifying unit 211 to the specified-block analyzing unit
11, the disk-block managing unit 21 is provided with the function as the file-system
managing unit.
[0053] The difference between the 3rd embodiment of the present invention and the 1st and
2nd embodiments thereof is as follows: The target blocks that the block-selection
specifying unit 211 selects as the criterion of the block selection are the blocks
including the data of the file/directory specified as the snapshot creation target
from the outside, and the blocks including the data related therewith.
[0054] Furthermore, the 3rd embodiment of the present invention permits the user to watch
the snapshot data as one physical volume. In order to make it possible for the user
to make reference to the snapshot data within this physical volume, it is required
to provide the physical volume in a manner of being able to be mounted as a file system.
In order to permit the physical volume including the snapshot data to be watched as
the mountable file system, it is required to reproduce the consistency of the file
system at the point-in-time of the snapshot creation without changing this consistency
at all.
[0055] FIG. 7 is a diagram for explaining a layout example of the file system in the disk
apparatus. FIG. 8 is a diagram for explaining an example of the logical structure
of the file system. Next, referring to FIG. 7 and FIG. 8, the explanation will be
given below concerning the file system in the 3rd embodiment.
[0056] The file system is configured as follows: As illustrated in FIG. 7, a volume (i.e.,
a set of disk blocks) allocated within the disk apparatus is divided into plural block
groups "0" to "N" 700 to 710. Moreover, the inside of each block group is divided
as illustrated by 720 to 770, thereby separating the usage forms. The same super block
720 and the same group descriptor 730 are held within each block group. If data of
the super block 720 in the block group "0" 700 is crashed, data of the block group
"1" or after is used. A data-block bit map 740 holds information on the usage situation
of a data block 770. An i-node bit map 750, which holds metadata based on file names
and block addresses at which data exists, holds information on the usage situation
of an i-node table 760 used for managing the i-node of a file/directory. The data
block 770 is used for a usage form in which the file data and directory entries are
stored.
[0057] In the example of the logical structure of the file system illustrated in FIG. 8,
the location of a root i-node 810 is held as a 1st entry of the i-node table 760.
The root i-node 810 refers to a data block 820 that stores the directory entries of
a root directory. The data block 820 stores therein a list that refers to names of
the files/directories existing in the directory and the i-node of each file/directory.
The i-node of each file refers to a data block (i.e., data block 830) that stores
the data. The i-node of each directory refers to a data block (i.e., data block 840)
that stores the directory entries.
[0058] Simultaneously, each i-node illustrated in FIG. 8 holds data that refers to an each
i-node's own parent directory. This makes it possible to search for, from the i-node
of a file or a directory, where the i-node's own parent directory exists, and further
where a directory above the parent directory exists. Namely, recognizing the i-node
of a file/directory makes it possible to search for a path from the root directory
to the file/directory.
[0059] At the time of mounting the file system, the super block 720 and the group descriptor
730 existing in the front-head block group 700 are read in, thereby making reference
to the management information on the file system. In order to make mountable the physical
volume including the snapshot data, in addition to the block including the metadata
and the data of the target file/directory, it is required to store the following information:
The super block 720 and the group descriptor 730 existing in the block group "0" 700,
and the data-block bit map 740 and the i-node bit map 750 within the block group to
which the block including the metadata and the data of the target file/directory belongs.
[0060] Also, after mounting the physical volume of the snapshot data, the 3rd embodiment
of the present invention, with the image that is the same as the one of the operational
data, permits the user to make reference to the path from the root directory to the
file/directory that has become the snapshot-data generation target. This makes it
necessary to store, in addition to the above-described file-system managing information,
the blocks that include all of the metadata and the directory entries of directories
existing on the path from the root directory to the target file/directory.
[0061] FIG. 9 is a flowchart for explaining the processing steps of the block selection
performed in the block-selection specifying unit 211 in the case where the snapshot-data
generation specified-target specified by the user is a file. Next, the explanation
will be given regarding this flowchart.
(1) The user specifies, as a snapshot-data generation target-block, a block including
a super block and a group descriptor of a file system to which the specified file
belongs (step 510).
(2) A block including a data-block bit map and an i-node bit map of a block group
to which the specified file belongs is specified as the snapshot-data generation target-block
(step 515).
(3) A block including metadata and data of the specified file is specified as the
snapshot-data generation target-block (step 520).
(4) A block including metadata and a directory entry of a parent directory of the
specified file is specified as the snapshot-data generation target-block (step 530).
(5) It is confirmed whether or not a further parent directory exists over the parent
directory of the specified file (step 540).
(6) If, in the judgement at the step 540, the further parent directory exists over
the parent directory, the processing goes back to the step 530, and basically the
same processing is continued with respect to this directory.
(7) If, in the judgement at the step 540, no directory exists over the parent directory,
the processing is terminated.
[0062] FIG. 10 is a flowchart for explaining the processing steps of the block selection
performed in the block-selection specifying unit 211 in the case where the snapshot-data
generation specified-target specified by the user is a directory. Next, the explanation
will be given regarding this flowchart.
(1) A block including a super block and a group descriptor of a file system to which
the directory specified by the user belongs is specified as a snapshot-data generation
target-block (step 600).
(2) A block including a data-block bit map and an i-node bit map of a block group
to which the specified directory belongs is specified as the snapshot-data generation
target-block (step 605).
(2) A block including metadata and a directory entry of the specified directory is
specified as the snapshot-data generation target-block (step 610).
(3) A block including metadata and data of all the files under the specified directory
is specified as the snapshot-data generation target-block (step 620).
(4) It is confirmed whether or not a parent directory of the specified directory exists
(step 630).
(4) If, in the judgement at the step 630, no parent directory exists, the processing
is terminated.
(5) If, in the judgement at the step 630, the parent directory exists, a block including
metadata and a directory entry of the parent directory is specified as the snapshot-data
generation target-block (step 640).
(6) It is confirmed whether or not a further parent directory exists over a directory
into which the block of the metadata and the directory entry has been stored at the
step 640 (step 650).
(7) If, in the judgement at the step 650, no directory exists over the directory further,
the processing is terminated.
(8) If, in the judgement at the step 650, the further parent directory exists over
the directory, the processing goes back to the step 640, and the processing is performed
again with this directory specified as the target.
[0063] Each processing according to each of the above-described embodiments of the present
invention can be configured as a processing program. This processing program can be
provided by being stored into a storage medium such as a HD, a DAT, a FD, a MO, a
DVD-ROM, or a CD-ROM.
[0064] The above-described 1st, 2nd, and 3rd embodiments can be co-used together. Also,
in the 1st, 2nd, and 3rd embodiments, the NAS processing unit and the disk apparatus
may be set up at locations that are away from each other, and a wide area network
may establish the connection therebetween. Otherwise, the NAS processing unit and
the disk apparatus may be configured in an integrated manner. In this case, the network
may be a system path for simply connecting both of them.
[0065] As having been explained so far, in the snapshot creation in the disk apparatus,
the present invention makes it possible to reduce the disk capacity needed for the
snapshot creation. Also, in the snapshot creation in the disk apparatus, the present
invention makes it possible to issue the snapshot creating instruction in the unit
of the partition or file/directory managed in the NAS processing unit.
[0066] It should be further understood by those skilled in the art that although the foregoing
description has been made on embodiments of the invention, the invention is not limited
thereto and various changes and modifications may be made without departing from the
spirit of the invention and the scope of the appended claims.
1. A storage system, comprising:
a 1st data storage (13),
a 2nd data storage (14) connected to said 1st data storage (13),
a processor (18) for controlling said 1st data storage (13) and said 2nd data storage
(14), and
data-block analyzing units (11, 21) for receiving data-block information so as to
store said data-block information into a storage region, said data-block information
indicating which data block is a useful data block out of a plurality of data blocks
in said 1st data storage (13).
2. The storage system as claimed in Claim 1, wherein said data-block analyzing unit (11)
has a plurality of bits (112) each corresponding to said plurality of data blocks
in said 1st data storage (13), said data-block analyzing unit (11) also setting a
bit (112) when receiving said data-block information, said bit (112) corresponding
to said data block among said plurality of bits (112).
3. The storage system as claimed in Claim 2, wherein, in correspondence with a writing-in
access to said 1st data storage (13), said processor (18) stores, into said 2nd data
storage (14), the copy of said data block to which said correspondence bit (112) has
been set.
4. The storage system as claimed in Claim 3, wherein, after having stored, into said
2nd data storage (14), said copy of said data block to which said correspondence bit
(112) had been set, said processor (18) generates a correspondence table (400) so
as to store said correspondence table (400) into said storage system, said correspondence
table (400) being a correspondence table between each address of said each copy-corresponding
data block in said 1st data storage (13) and each address of each copy-corresponding
data block in said 2nd data storage (14).
5. The storage system as claimed in Claim 1, further comprising a block-selection specifying
unit (211) connected among a network, said 1st data storage (13), said 2nd data storage
(14), and said data-block analyzing unit (21), said block-selection specifying unit
(211) also generating and managing said data-block information.
6. The storage system as claimed in Claim 5, wherein said data-block analyzing unit (21)
has a plurality of bits (112) each corresponding to said plurality of data blocks
in said 1st data storage (13), said data-block analyzing unit (21) also setting a
bit (112) when receiving said data-block information, said bit (112) corresponding
to said data block among said plurality of bits (112).
7. The storage system as claimed in Claim 5, wherein, in correspondence with a writing-in
access to said 1st data storage (13), said processor (18) stores, into said 2nd data
storage (14), the copy of said data block to which said correspondence bit (112) has
been set.
8. A snapshot creating method in a storage system including a file server (2) and a disk
apparatus (1), said snapshot creating method comprising:
a step of instructing, from said file server (2) to said disk apparatus (1), a block
whose snapshot data should be generated, and
a step (300) of, in said disk apparatus (1), if there occurs an updating request for
operational data, confirming whether or not an updated-data written-in destination
block in an operational data storing region (13) has become a snapshot-data creation
specified block, and
a step (360) of, if said updated-data written-in destination block has not become
a target of said snapshot data generation, instructing a disk controller (12) to write
updated data into said operational data storing region (13), and
a step (310) of, if said updated-data written-in destination block has become said
snapshot-data generation specified block, confirming whether or not said specified
block had been copied into a snapshot data storing region (14) before, and
said step (360) of instructing said disk controller (12), if said specified block
had been copied before, to write said updated data into said operational data storing
region (13), and
a step (320) of, if said specified block had not been copied before, instructing said
disk controller (12) to copy pre-updated data into said snapshot data storing region
(14), said pre-updated data existing in said operational data storing region (13),
and
a step (330) of receiving said instruction of copying said pre-updated data, and confirming
whether or not said snapshot data storing region (14) has been ensured, and
a step (340) of, if said snapshot data storing region (14) has not been ensured, ensuring
said snapshot data storing region (14), and
a step (350) of copying said pre-updated data into said snapshot data storing region
(14), and generating simultaneously a block-address correspondence table (400) if
necessary, and
said step (360) of writing said updated data into said written-in destination block
in said operational data storing region (13).
9. The snapshot creating method as claimed in Claim 8, wherein said file server (2) has
in advance written, as a bit map, said instruction of said block into a determined
storing region within said disk apparatus (1).
10. The snapshot creating method as claimed in Claim 8, wherein said file server (2) transmits
said updated data to said disk apparatus (1), and at said same time transmits said
instruction of said block as additional information.
11. The snapshot creating method as claimed in Claim 8, wherein, when said disk apparatus
(1) generates said snapshot data, said disk apparatus (1) creates said instruction
of said block by making an inquiry thereabout to said file server (2).
12. The snapshot creating method as claimed in Claim 8, wherein said instruction of said
block instructs a block included in a specific partition within said disk apparatus
(1).
13. The snapshot creating method as claimed in Claim 8, wherein said instruction of said
block is performed under a condition of excluding a block that said file server (2)
does not use, or a block that said file server (2) manages as a free block which is
an ineffective block.
14. The snapshot creating method as claimed in Claim 8, wherein said instruction of said
block is for said file server (2) to instruct a block including metadata and data
of a specific file, a block including metadata that becomes necessary for mounting
a file system, and a block including metadata and data of a directory that exists
on a path from a root directory to said specific file.
15. A disk apparatus (1), comprising:
specified-block analyzing means (11) for receiving, from a file server (2), an instruction
of a block specified as target of a snapshot creation, and for making a judgement
as to whether the snapshot data generation is of a necessity or not in accordance
with said instruction,
a disk controller (12) for executing copy of said data on the basis of said judgement
by said specified-block analyzing means (11),
an operational data storing region (13) for saving operational data, and
a snapshot data storing region (14) for saving said snapshot data.
16. The disk apparatus (1) as claimed in Claim 15, wherein said specified-block analyzing
means (11) includes a block bit map storing region (112) for storing said information
on said block instruction received from said file server (2).
17. A storage system, comprising:
a file server (2), and
a disk apparatus (1), wherein said file server (2) includes means (21) for instructing
said disk apparatus (1) in a block whose snapshot data should be generated.
18. The storage system as claimed in Claim 17, wherein said file server (2) includes means
(211) for transmitting updated data to said disk apparatus (1), and at the same time
transmitting said instruction of said block as additional information.
19. A storage system, comprising:
a file server (2), and
a disk apparatus (1), wherein said disk apparatus (1) includes means that, when said
disk apparatus (1) generates snapshot data, makes an inquiry to said file server (2),
said inquiry being about a block whose snapshot data should be generated.
20. A computer-readable processing program for creating a snapshot in a disk apparatus
(1), comprising:
a program (11') for, if there occurs an updating request for operational data, confirming
whether or not an updated-data written-in destination block in an operational data
storing region (13) has become a snapshot-data creation specified block, and
said program (11') for, if said updated-data written-in destination block has not
become a target of said snapshot data generation, instructing a disk controller (12)
to write updated data into said operational data storing region (13), and
said program (11') for, if said updated-data written-in destination block has become
said snapshot-data generation specified block, confirming whether or not said specified
block had been copied into a snapshot data storing region (14) before, and
said program (11') for, if said specified block had been copied before, instructing
said disk controller (12) to write said updated data into said operational data storing
region (13), and
said program (11') for, if said specified block had not been copied before, instructing
said disk controller (12) to copy pre-updated data into said snapshot data storing
region (14), said pre-updated data existing in said operational data storing region
(13), and
a program (12') for receiving said instruction of copying said pre-updated data, and
confirming whether or not said snapshot data storing region (14) has been ensured,
and
said program (12') for, if said snapshot data storing region (14) has not been ensured,
ensuring said snapshot data storing region (14), and
said program (12') for copying said pre-updated data into said snapshot data storing
region (14), and generating simultaneously a block-address correspondence table (400)
if necessary, and
said program (12') for writing said updated data into said written-in destination
block in said operational data storing region (13).